Vedecký časopis FINANČNÉ TRHY, Bratislava, Derivat 2018, ISSN 1336-5711, 1/2018

Impact of Regulatory Stress Tests on Banking Sectors in the EU and US1

Lukáš MAJER

Abstract

Simulating the conditions of economic recession and estimating its impact on bank risks

and performance can help banks to prepare its portfolios for negative consequences of financial

shocks. In this context regulators worldwide conducted series of system wide stress tests with the

motivation to assess the financial institution health. The primary goal of this article is to analyze

the significance of regulatory system wide stress tests impact on individual bank institutions and

the banking sector as a whole. The analysis was performed in two phases. At first, to identify

short term impact, we analyzed changes in the bank share price volatility, during the period of

stress test results announcement. In the second phase, we constructed a linear autoregressive

model (AR2) with structural dummy variables representing the stress test results announcement

shocks. Consequently, we assessed the shock variables’ significance and its direction of

influence on bank shares in long run. The analysis was performed in two different stress testing

environments carried out by the European Banking Authority (EBA) and the US Federal Reserve

System (FED), and covered two system wide stress tests within the period of 2012 – 2017. We

see the main contribution of the paper in assessment of regulatory stress test impact on the EU

and US banking sector. Performed analysis was primarily based on market sensitivity to external

shocks assuming market efficiency to enable the evaluation of banking institutions’ health and

the sustainability of their businesses. The output of the analysis indicates, that the release of

system wide stress test results had a significant impact on banking sector performance, especially

in the US market.

Keywords: risk management, stress testing, volatility, EBA, FED

JEL classification: G31, E58

1. Introduction

As a reaction on financial crisis regulators worldwide conducted series of system wide stress

tests with the goal to assess the financial institution health and its resilience to withstand possible

disturbances in financial markets during economic recession. The primary goal of this article is

to analyze the significance of regulatory system wide stress tests impact on individual bank

institutions and banking sector itself. The analysis was done in two steps. Firstly, we analyzed

changes in bank share price volatility around stress test results announcements. Secondly, we

constructed a linear autoregressive model AR (2) with dummy variables representing stress test

results announcements, where we assess significance and direction of dummy variable

coefficients. The analysis is done on the sample of EU and US bank houses covering two system

wide stress tests within the 2012 – 2017 period. The sample consists of 24 banks with two types

1 This paper is an outcome of research project VEGA 1/0946/17 “Credit Cycle and its determinants in countries of

Central and Eastern Europe.”

of institutions. On the one hand, there are banks which failed to pass minimal 8% Core Equity

Tier 1 (CET1) capital requirements in severe stress test scenario. On the other hand, there are

benchmark institutions with the strong and resilient capital structure which CET1 capital

adequacy in severe scenario remained over 8 % threshold. We assumed that institutions who

failed stress test, would experience significant deterioration of marked value represented by

share price drop in period after stress test results announcement.

2. Theoretical Framework and Stress Test Environment Definition

According to Mishina et al. (2006) the ultimate objective of stress tests is to assess the

performance of the financial system under abnormal operating conditions. The financial system,

however, is a complex entity consisting of a wide range of financial institutions, financial

markets, and payments and settlement systems, and it is not easily amenable to aggregate

analysis. In practice, the analysis of financial system stability focuses on individual components,

most often the financial institutions, to arrive at an overall assessment of the financial system.

Moreover Virolainen (2004) points out that stress tests are generally used to complement

financial institutions’ internal models, such as value-at-risk (VaR). Standard VaR models have

been found to be of limited use in measuring financial institutions’ exposures to extreme market

events, i.e. events that occur too rarely to be captured by statistical models, which are normally

based on relatively short periods of historical data. In addition to applying stress tests to the

portfolios of individual institutions at the micro level, stress testing plays an important role in the

macro-prudential analysis of public authorities. The role of macro-prudential or financial

stability, analysis has gained in importance in recent years, both among central banks, regulatory

authorities and international agencies.

Decisions made by regulatory authorities or central banks, may have a significant and long

lasting impact on the entire banking sector. Borio et al. (2015) investigate the effects of

regulatory monetary policy thought bank profitability, using data of 109 large international banks

for the period 1995–2012. Author points on the fact, that there exists a certain relationship

between the interest rate structure and the bank profitability (return on assets). The paper

suggests that, over time, unusually low interest rates and an unusually flat term structure erode

bank profitability. Altunbasa et al. (2014) investigate the effect of relatively loose monetary

policy on bank risk through a large panel regression including quarterly information from listed

banks operating in the European Union and United States. Article suggest that relatively low

levels of interest rates over an extended period of time contributed to an increase in bank risk.

This result holds for a wide range of measures of risk, as well as macroeconomic and

institutional controls including the intensity of supervision, securitization activity, and bank

competition. The results suggest that monetary policy is not neutral from bank financial stability

perspective.

In volatility analysis, we focused on the impact of stress test results announcement on the bank

share prices volatility, which represents market participant’s opinion of bank financial health and

condition. In similar and more general context, Bomfim (2003) examines regulatory pre-

announcement and news effects on the stock market in the context of public disclosure of

monetary policy decisions. Author conclusions suggest that on days preceding regularly

scheduled policy announcements the stock market tends to be relatively quiet, with conditional

volatility abnormally low. The paper also analysis how the actual interest rate decisions of policy

makers affect the stock market volatility. The element of surprise in such decisions tends to boost

stock market volatility significantly in the short run, and positive surprises, higher-than-expected

values of the target federal funds rate tend to have a larger effect on volatility than negative

surprises. Very interesting analysis concerning announcement effects can be found in Li et al.

(2016) paper, where the authors study the differences in the announcement effects of seasoned

equity offerings (SEOs) of commercial banks and non-bank companies. Article explores the

influence of bank regulation and the financial crisis using propensity score matching-based

(difference-in-difference) analysis. Abnormal stock returns on SEO announcements for US

commercial banks are significantly higher than those of non-bank companies, which suggests,

that bank regulations reduce the likelihood that bank SEOs signal overpriced equity. Wang and

More (2007) investigate sudden changes in volatility in the stock markets of new EU members

using GARCH models to detect the sudden change in variance of returns and the length of this

variance shift. A sudden change in volatility seems to arise from the evolution of emerging stock

markets, exchange rate policy changes and financial crises. According to Bacchiocchi and

Fanelli (2014) a growing line of research makes use of structural changes and different volatility

regimes found in the data in a constructive manner to improve the identification of structural

parameters in the Structural Vector Autoregressions (SVARs). Authors approach generalizes the

existing literature on ‘identification through changes in volatility’ to a broader framework and

their empirical illustration focuses on a small monetary policy SVAR model of the U.S. economy

which suggests that monetary policy has become more effective at stabilizing the economy since

the 1980s.

2.1. US FED and EBA Stress Test Environment Definition

The Federal Reserve (2015, 2016) expects large, complex bank holding companies to have

sufficient capital to continue lending to support real economic activity while meeting their

financial obligations, even under stressful economic conditions. Stress testing is the tool that

helps bank supervisors to measure whether banks have enough capital to support its operations

throughout periods of stress. In 2015 the FED stress test was carried on a sample of 31 banks,

while in 2016 it was 33 banks. The models are based on hypothetical, stressful macroeconomic

and financial market scenarios developed by the Federal Reserve. The severely adverse scenario

features a deep recession in the United States, Europe, and Japan, significant declines in asset

prices and increases in risk premia, and a marked economic slowdown in developing Asia. In

2016 severely adverse scenario is characterized by a severe global recession accompanied by a

period of heightened corporate financial stress and negative yields for short-term U.S. Treasury

securities.

According to EBA (2014, 2016) the objective of the EU‐wide stress test is to assess the

resilience of banks in the EU to adverse economic developments, helping supervisors assess

individual banks, contributing to understanding systemic risk in the EU and fostering market

discipline. The 2014 stress test includes 123 banking groups across the EU and including

Norway with a total of EUR 28,000BN of assets covering more than 70% of total EU banking

assets. The 2016 sample was reduced to 55 EU banks. The EU‐wide stress test is coordinated by

the EBA across the EU. The proposed adverse scenario reflects the systemic risks representing

threats to banking sector stability such as an increase in global bond yields, especially in

emerging market economies. Further deterioration of credit quality in countries with feeble

demand, weak fundamentals and still vulnerable banking sectors. Stalling policy reforms

jeopardizing confidence in the sustainability of public finances and the lack of necessary bank

balance sheet repair to maintain affordable market funding.

2.2. EBA and FED Stress Test Results Comparison

In tables below, we analyze EU and US stress test results. The analysis is based on counts of

entities whose CET1 ratio decreased under 8%, 6%, 5% and finally 0% threshold. We compare

EU and US results separately.

Table 1 Stress Test Adverse Scenario Results

CET1 Band Total Adverse Adverse % Total Adverse Adverse % CET1 Band Total Adverse Adverse % Total Adverse Adverse %

< 8% 123 51 41% 51 9 18% < 8% 31 12 39% 33 14 42%

< 6% 123 25 20% 51 1 2% < 6% 31 1 3% 33 2 6%

< 5% 123 18 15% 51 1 2% < 5% 31 0 0% 33 0 0%

< 0% 123 5 4% 51 1 2% < 0% 31 0 0% 33 0 0%

Average Average

Median Median9.5%

EU

8.1% 8.3%

US

9.2% 10.7% 9.2% 8.9%

2015 2016

8.2%

2014 2016

Source: Author calculation according to EBA (2014, 2016) and FED (2014, 2016) data

The table above displays stress test results in two dimensions, adverse CET1 evolution between

two stress test periods and comparison of EU and US banks which failed to meet CET1 ratio

thresholds. Analyzing the simple average and median figures there has been a certain

improvement in the case of EU where the median CET1 ratio in adverse scenario increased from

8.2% to 9.5%. In the US, the growth was just 0.2% while average figure decreased by 0.3%. In

2014 EU stress test 41% of the sample reported adverse CET1 ratio below 8%, and 20% even

below 6%, with 5 banks reporting negative figures. In 2016, there has been a significant

improvement where only 18 % of the sample experienced CET1 ratio drop below 8%. In the case

of US stress test these figures remained almost unchanged showing approximately 40 % of

institutions with adverse CET1 ratio lower than 8%, while no bank reported ratios below 5%.

2.3. Representative Sample of EU and US Banking Institutions and Data Definition

To analyze the stress test impact a representative sample of 24 banks was selected with the goal

to cover as much variability as possible. Constructing the sample, we took into the consideration

the scope of the analysis as well as restrictions in bank shares availability, where not all bank

shares are publicly traded or the market liquidity is low. The selected sample is divided in two

dimensions. At first, we divide banks according to its origin where in the EU we analyze 16 and

in the US 8 institutions. Further on we split the sample by the results in 2014/2015 adverse stress

scenario on “BAD” institutions where the adverse CET1 ratio < 8% and “GOOD” ones where

the adverse CET1 ratio > 8%. The entire sample overview is summarized in Table2.

Table 2 Representative Sample of EU and US Banking Institutions

NAME CODE EU/US COUNTRY ASOF ADVERSE GOOD/BAD ASOF ADVERSE GOOD/BAD CHANGE

Erste Group Bank AG EBKDY EU AUSTRIA 10.0% 7.6% BAD 12.4% 8.2% GOOD YES

Bank of Cyprus Public Company Limited BOC.AT EU CYPRUS 7.3% 1.5% BAD n/a n/a n/a n/a

Piraeus Bank S.A. TPEIR.AT EU GREECE 10.0% 4.4% BAD n/a n/a n/a n/a

Allied Irish Banks, p.l.c. AIBSF EU IRELAND 14.6% 6.9% BAD 15.9% 7.4% BAD NO

Banco Comercial Português S.A. BPCGF EU PORTUGAL 10.3% 3.0% BAD n/a n/a n/a n/a

LIBERBANK LBK.MC EU SPAIN 7.8% 5.6% BAD n/a n/a n/a n/a

The Royal Bank of Scotland Group plc RBS EU UK 8.6% 5.7% BAD 15.5% 8.1% GOOD YES

Barclays PLC BCS EU UK 9.1% 7.1% BAD 11.4% 7.3% BAD NO

KBC Group NV KBCSY EU BELGIUM 12.7% 8.3% GOOD 15.2% 11.3% GOOD NO

DANSKE BANK DSN.BE EU DENMARK 13.7% 11.7% GOOD 16.1% 14.0% GOOD NO

Societe Generale Group GLE.PA EU FRANCE 10.7% 8.1% GOOD 11.4% 8.0% GOOD NO

Deutsche Bank AG DB EU GERMANY 13.4% 8.9% GOOD 14.7% 9.9% GOOD NO

OTP BANK OTP.F EU HUNGARY 15.9% 11.9% GOOD 13.4% 9.2% GOOD NO

Intesa Sanpaolo S.p.A. ISP.MI EU ITALY 11.7% 8.3% GOOD 13.0% 10.2% GOOD NO

BANCO SANTANDER SAN.MC EU SPAIN 10.4% 8.9% GOOD 12.7% 8.7% GOOD NO

HSBC Holdings plc HSBC EU UK 10.8% 9.3% GOOD 11.9% 8.8% GOOD NO

The Goldman Sachs Group, Inc. GS USA US 15.1% 5.8% BAD 13.6% 8.4% GOOD YES

Zions Bancorporation ZION USA US 11.9% 6.0% BAD 12.2% 6.6% BAD NO

Morgan Stanley MS USA US 15.2% 6.3% BAD 16.4% 9.1% GOOD YES

Citigroup Inc. C USA US 15.1% 6.8% BAD 15.3% 9.2% GOOD YES

Regions Financial Corporation RF USA US 11.8% 8.5% GOOD 10.9% 7.3% BAD YES

Huntington Bancshares Incorporated HBAN USA US 10.3% 8.7% GOOD 9.8% 5.0% BAD YES

American Express Company AXP USA US 13.6% 13.0% GOOD 12.4% 11.4% GOOD NO

State Street Corporation STT USA US 13.9% 8.1% GOOD 13.0% 9.6% GOOD NO

US

GO

OD

2014/2015 2016EU

BA

DEU

GO

OD

US

BA

D

Source: EBA (2014, 2016) and FED (2015, 2016) data

The daily share prices data was downloaded from finance.yahoo.com covering the period from

1.1.2012 to 10.1.2017. In the case of EU institutions, we are interested in share price reaction

after stress test result announcement in October 2014 (26/10/2014) and July 2016 (29/07/2016)

while in case of US institutions it is March 2015 (11/03/2015) and June 2016 (29.6.2016).

3. Stress Test Analysis Results

In the analytical part of the paper, we firstly perform correlation analysis to see dependencies

between bank institutions within different economic environments, as well as we test correlations

between “BAD” and “GOOD” performing banks. Further on we analyze the bank share volatility

changes with the goal to assess immediate impact of results announcement on bank shares. At

last we present results of AR(2) regression models where we are interested in long term impact,

analyzing structural (dummy) variables significance and the direction of influence.

3.1. Correlation Analysis Results

Analyzing correlations between bank shares within the sample, we can observe strong positive

correlations across the whole US sector, while the EU market is more heterogenous. Analysis

shows also certain correlations between EU and US institutions, however the relationship is

weaker then within EU and US sectors it selves. Certain positive correlations between banking

institutions could have contributed to significant impact of 2016 results announcements,

especially across the US banking sector.

Table 3 Correlation Analysis

EBKDY BOC.AT TPEIR.AT AIBSF BPCGF LBK.MC RBS BCS KBCSY DSN.BE GLE.PA DB OTP.F ISP.MI SAN.MC HSBC GS ZION MS C RF HBAN AXP STT

EBKDY 100%

BOC.AT 22% 100%

TPEIR.AT 44% 72% 100%

AIBSF -24% -83% -56% 100%

BPCGF 18% -27% -2% 13% 100%

LBK.MC 2% -55% -24% 55% 17% 100%

RBS 19% 88% 65% -86% -31% -38% 100%

BCS 45% 86% 61% -88% -13% -48% 86% 100%

KBCSY 15% -10% -22% -6% 18% 32% 5% -4% 100%

DSN.BE -35% -59% -73% 42% 22% 39% -46% -58% 71% 100%

GLE.PA 43% 29% 22% -45% 29% 15% 42% 41% 76% 26% 100%

DB 56% 86% 79% -78% -13% -43% 80% 93% -20% -77% 33% 100%

OTP.F -9% -81% -74% 76% 42% 40% -87% -70% 22% 64% -13% -75% 100%

ISP.MI -16% -8% -40% -12% -11% 27% 16% -2% 81% 73% 60% -26% 10% 100%

SAN.MC 17% 85% 72% -80% -5% -33% 87% 73% 16% -34% 53% 71% -74% 14% 100%

HSBC 42% 85% 81% -80% 1% -48% 78% 86% -22% -73% 27% 91% -72% -32% 80% 100%

GS -11% -1% -27% -22% 49% 0% 9% 10% 59% 55% 58% -14% 24% 55% 23% 2% 100%

ZION 27% -6% 2% -6% 78% -4% -17% 6% 32% 23% 44% -1% 37% 0% 10% 17% 69% 100%

MS -13% 9% -13% -31% 43% -3% 18% 12% 66% 54% 63% -10% 10% 57% 38% 9% 94% 64% 100%

C 13% 31% 1% -51% 37% -23% 36% 47% 41% 18% 60% 25% -3% 35% 43% 39% 86% 68% 82% 100%

RF 22% 1% 12% -13% 79% 2% -6% 10% 35% 21% 50% 3% 25% 4% 24% 22% 70% 96% 69% 68% 100%

HBAN -15% -29% -45% 6% 45% 26% -15% -19% 75% 78% 60% -40% 44% 71% 0% -32% 88% 60% 85% 65% 62% 100%

AXP 21% 80% 76% -77% 6% -34% 80% 69% 14% -36% 50% 69% -73% 4% 92% 78% 25% 23% 41% 46% 36% 4% 100%

STT 2% 34% 5% -50% 33% -29% 37% 39% 44% 21% 53% 19% -8% 33% 50% 39% 83% 62% 88% 90% 64% 63% 53% 100%

US

BA

DU

S G

OO

D

EU BAD EU GOOD US BAD US GOOD

EU B

AD

EU G

OO

D

Source: Author calculation based on finance.yahoo.com data

Understanding correlations between institution shares within our sample is useful for further

interpretations of the short term as well as the long term stress test impact within the volatility

and regression analysis.

3.2. Volatility Analysis Results

Analyzing bank share volatility changes in the period around stress test results announcement,

we can observe significantly higher volatility in 2016, specially across the whole US sector. On

the one hand, this increased volatility proves that there is a significant, immediate impact of

stress result announcement on bank shares. On the other hand, our assumption that institution

whose CET1 ratio in stress decreased below 8% will experience drop in the share value was not

confirmed. In contrary we can see a positive market reaction across the whole industry. In the

case of EU 2014 and US 2015 stress test, there seems to be no or only limited volatility changes

around observed period, which indicates no or only minor short term stress test impact.

Figure 1 Changes in Historical Volatility of Bank Shares

Source: Author calculation based on finance.yahoo.com data

Charts above show the historic volatility of selected bank institutions. The entire volatility

analysis for whole sample can be found in Appendix 1. Analysis of volatility changes helped us

to understand immediate, short term impact of stress results announcement on bank performance.

To identify the long term, structural impact of stress tests we constructed the autoregressive

model with dummy variables representing stress test results announcement. From the analysis, it

is also obvious, that the US results had quite significant impact on the EU institutions as well.

3.3. Regression Analysis Results

In following section, charts displays the evolution of share prices of EU and US banks, whereas

tables contain results of AR (2) regression models where dummy variables EU_2014, EU_2016,

US_2015 and US_2016 represents shocks caused by the stress test results announcement. In

analysis, we are interested in the significance of dummy variable coefficients and its direction of

influence.

3.3.1. EU and US “BAD” Institutions Results

Bad institutions are defined as the one, whose CET1 capital adequacy ratio in the adverse stress

scenario decreased below 8%. For these institutions, we would expect that negative message

about its capital situation will have a negative impact on share prices. Charts and tables below

show the impact of two stress test shocks, the first in 2014 (US 2015) and second in 2016.

Comparing adverse CET1 capital ratios between two periods, we can observe an improvement of

capital adequacy in case of all EU and US banks within the sample. In the case of EU there are 4

banks, mainly from southern Europe which were not included in 2016 exercise.

Figure 1 and Figure 2 display the bank share prices time series as well as results of AR(2) model,

where we focus on impact of stress test dummy variables on EU and US “BAD” banks.

Figure 2 Share Prices Evolution in Time and Regression Results for EU BAD Banks

0

0.5

1

1.5

2

2.5

3

3.5

-4

-3

-2

-1

0

1

2

3

01/02/12

04/02/12

07/02/12

10/02/12

01/02/13

04/02/13

07/02/13

10/02/13

01/02/14

04/02/14

07/02/14

10/02/14

01/02/15

04/02/15

07/02/15

10/02/15

01/02/16

04/02/16

07/02/16

10/02/16

01/02/17

BOC.AT - 1.5 % |n/a TPEIR.AT - 4.4 % |n/a AIBSF - 6.9 % |7.39 % BPCGF - 3 % |n/a

LBK.MC - 5.6 % |n/a EBKDY - 7.6 % |8.19 % RBS - 5.7 % |8.08 % BCS - 7.1 % |7.3 %

21

EU BAD Banks

TPEIR.AT

RBS

BCS

BPCGF

Adverse CET1 2014 / 2016 (Intercept) lag_1 lag_2 eu_2014 eu_2016

EBKDY 7.6 % - 8.19 % ** | + ** | +

BOC.AT 1.5 % - n/a ** | + ** | +

TPEIR.AT 4.4 % - n/a ** | + ** | + * | - ** | -

AIBSF 6.9 % - 7.39 % ** | +

BPCGF 3.0 % - n/a ** | + ** | +

LBK.MC 5.6 % - n/a ** | + ** | + * | -

RBS 5.7 % - 8.08 % ** | + ** | + * | -

BCS 7.1 % - 7.3 % ** | + ** | + ** | - * | -

*0.1 | **0.05 | + positive impact | - negative impact Source: Author calculation in R based on finance.yahoo.com data

Figure 3 Share Prices Evolution in Time and Regression Results for US BAD Banks

2.2

2.7

3.2

3.7

4.2

2.8

3.3

3.8

4.3

4.8

5.3

01

/02

/12

04

/02

/12

07

/02

/12

10

/02

/12

01

/02

/13

04

/02

/13

07

/02

/13

10

/02

/13

01

/02

/14

04

/02

/14

07

/02

/14

10

/02

/14

01

/02

/15

04

/02

/15

07

/02

/15

10

/02

/15

01

/02

/16

04

/02

/16

07

/02

/16

10

/02

/16

01

/02

/17

GS - 5.8 % |8.4 % ZION - 6 % |6.6 % MS - 6.3 % |9.1 % C - 6.8 % |9.2 %

1 2

Adverse CET1 2015 / 2016 (Intercept) lag_1 lag_2 us_2015 us_20016

GS 5.8 % - 8.4 % ** | + ** | +

ZION 6 .0% - 6.6 % ** | + ** | + ** | +

MS 6.3 % - 9.1 % ** | + ** | +

C 6.8 % - 9.2 % ** | + ** | + ** | +

*0.1 | **0.05 | + positive impact | - negative impact Source: Author calculation in R based on finance.yahoo.com data

As already mentioned for “BAD” institutions we would expect a significant deterioration of

share prices after the stress test results announcement. From figures above it is obvious, that in

most cases, there seems to be only limited impact on bank shares where only 3 banks in Europe

proved significant coefficients of the 2014 shock. Taking into an account results of the volatility

analysis, the example of TPIER.AT, RBS and BCS, shows rather long term influence, which

contributed to bank shares value deterioration. If we consider results of US 2015 exercise, it

seems to have no significant impact at all.

Having a look at 2016 results there seems to be a relatively strong positive impact of stress test

results across the whole sector in US. This could be driven by the fact, that all institutions

improved its capital situation as well as presence of the strong correlation between whole

banking sector in the US. The EU 2016 stress test seems to be insignificant with only exception

of BPCGF bank. Similarly, to volatility analysis, regression results also prove that there is a

minimal sensitivity of markets towards the fact that banks CET1 ratio in adverse scenario

decreased below psychological threshold of 8%.

US BAD Banks

3.3.2. EU and US “GOOD” Institutions Analysis

Conversly to BAD institutions a GOOD one are defined as those which CET1 capital adequacy

ratio in adverse stress scenario remained over 8% threshold. For these institutions, we would

expect that a good assessment of its capital position will have positive or at least not negative

effect on share prices.

Figure 3 and figure 4 show the bank share prices time series as well as results of AR(2) model

where we focus on impact of stress test dummy variables on EU and US “GOOD” banks.

Figure 4 Share Prices Evolution in Time and Regression Results for EU GOOD Banks

1

1.5

2

2.5

3

3.5

4

4.5

-0.1

0.4

0.9

1.4

1.9

2.4

2.9

3.4

3.9

01

/02

/12

04

/02

/12

07

/02

/12

10

/02

/12

01

/02

/13

04

/02

/13

07

/02

/13

10

/02

/13

01

/02

/14

04

/02

/14

07

/02

/14

10

/02

/14

01

/02

/15

04

/02

/15

07

/02

/15

10

/02

/15

01

/02

/16

04

/02

/16

07

/02

/16

10

/02

/16

01

/02

/17

DSN.BE - 11.7 % |14.02 % GLE.PA - 8.1 % |8.03 % DB - 8.9 % |9.85 % OTP.F - 11.9 % |9.22 %

ISP.MI - 8.3 % |10.24 % KBCSY - 8.3 % |11.27 % SAN.MC - 8.9 % |8.69 % HSBC - 9.3 % |8.76 %

1 2

Adverse CET1 2014 / 2016 (Intercept) lag_1 lag_2 eu_2014 eu_2016

KBCSY 8.3 % - 11.27 % ** | + ** | +

DSN.BE 11.7 % - 14.02 % ** | + ** | + * | +

GLE.PA 8.1 % - 8.03 % ** | + ** | + * | +

DB 8.9 % - 9.85 % ** | +

OTP.F 11.9 % - 9.22 % ** | + ** | + ** | + ** | + ** | +

ISP.MI 8.3 % - 10.24 % ** | + ** | +

SAN.MC 8.9 % - 8.69 % * | + ** | + * | -

HSBC 9.3 % - 8.76 % ** | + ** | + ** | -

*0.1 | **0.05 | + positive impact | - negative impact Source: Author calculation in R based on finance.yahoo.com data

Figure 5 Share Prices Evolution in Time and Regression Results for US GOOD Banks

EU GOOD Banks

DSN.BE

GLE.PA

OTP.F

HSBC

SAN.MC

2

2.5

3

3.5

4

4.5

1.4

1.6

1.8

2

2.2

2.4

2.6

2.8

01

/02

/12

04

/02

/12

07

/02

/12

10

/02

/12

01

/02

/13

04

/02

/13

07

/02

/13

10

/02

/13

01

/02

/14

04

/02

/14

07

/02

/14

10

/02

/14

01

/02

/15

04

/02

/15

07

/02

/15

10

/02

/15

01

/02

/16

04

/02

/16

07

/02

/16

10

/02

/16

01

/02

/17

RF - 8.5 % |7.3 % HBAN - 8.7 % |5 % AXP - 13 % |11.4 % STT - 8.1 % |9.6 %

1 2

Adverse CET1 2015 / 2016 (Intercept) lag_1 lag_2 us_2015 us_20016

RF 8.5 % - 7.3 % ** | + ** | + ** | +

HBAN 8.7 % - 5 % ** | + ** | +

AXP 13 % - 11.4 % ** | + ** | + * | +

STT 8.1 % - 9.6 % ** | + ** | + ** | +

*0.1 | **0.05 | + positive impact | - negative impact Source: Author calculation in R based on finance.yahoo.com data

Analyzing 2016 results there seems to be a relatively strong positive impact of stress test results

across the whole US sector. In case of EU 2016 stress test there are banking institutions where

the impact was positive. Contra-intuitive results can be observed in the EU 2014 case of HSBC

and SAN.MC for which we identified significant negative impact despite the fact that its capital

position in stress remained strong.

4. Discussion and conclusions

Banking in general can be considered as one of the fastest growing and most regulated industries.

Under the pressure of last financial crisis, the entire banking system is undergoing major changes

that require adequate and timely response from banking institutions as well as regulators. It

should be the primary interest of banking institutions to maintain sound internal processes and

try to predict and quickly react to sudden changes and threats from external environment. This

should consequently help them to mitigate negative effects of economic cycles. The analysis of

two different banking environments helped us to understand the impact of stress test result in

slightly broader context, where we could observe certain correlations between EU and US

institutions. From 2016 volatility analysis, it seems that US results had quite significant impact

on EU institutions as well.

The analysis has proven that results of regulatory system wide stress tests have its importance

and should not be underestimated. At the same time, it is important to realize that any model is

just a simplification of the reality, and model results should be considered just like the general

guideline and indication of further possible directions. Positive outcomes of 2016 stress test

should not stop banking institutions from maintaining sound risk profiles, as well as strong

capital structure, which are cornerstones of sustainable banking business and stability of entire

financial sector.

US GOOD Banks

For the further research in this area, it would be interesting to use more advanced approaches for

the volatility analysis such as GARCH models, to detect the sudden change in variance of returns

and the length of this variance shift and associate these changes with stress test results

announcement shocks. Structural Vector Autoregressions (SVARs) models also provide an

interesting approach for volatility changes identification, which is more calculation and

methodology intensive. Moreover, it would be interesting to analyze changes in the banks’

capital and Risk Weighted Assets structure after 2014 period and understand what was the

banks’ reaction concerning changes in the risk profile and balance sheet.

Adequate stress test framework should be an essential part of the sound and complex risk

management strategy of each bank. The primary goal of this article was to analyze the impact of

regulatory stress tests results on banking institutions as well as entire banking sector. The

immediate, short term stress test impact was analyzed by changes in the bank’s share price

volatility around stress test results announcements. To identify potential long term influence we

constructed a linear autoregressive model (AR2) with dummy variables representing stress test

results announcement shocks. The outputs of both analyses indicate, that in the short, as well as

long term horizon, stress test results had a certain significant effect on bank performance.

In case of FED US stress test, 2015 exercise seemed to have no, or only very limited influence.

Significantly different situation was observed in 2016 where in all cases we identified relatively

strong positive impact on the whole sector. In case of EBA EU stress test the conclusion is not so

straightforward. The results are more heterogenous which can be driven by higher heterogeneity

within EU banking sector proved by correlation analysis. In 2014 stress test three banks with

insufficient capital in adverse scenario experienced negative impact of results announcement on

its market value. At the same time, we could observe contra-intuitive results for HSBC and

SAN.MC where we identified significant negative impact despite the fact that its capital position

in stress remained strong. In EU 2016 stress test there are four banking institutions where impact

was positive. Both volatility as well as regression analysis proved that there is only a minimal

sensitivity of markets towards the fact that banks CET1 ratio in adverse scenario decreased

below psychological threshold of 8%. Due to the certain correlations within banking sector,

especially within US market, it seems that the stress test impact is primarily driven by results of

the whole sector rather than individual institutions.

References

[1] ALTUNBASA, Y., GAMBACORTAB, L., MARQUES-IBANEZ, D. 2014. Does Monetary

Policy Affect Bank Risk?. International Journal of Central Banking, volume 10 (No. 1), 95 -

135. Retrieved from http://www.ijcb.org/journal/ijcb14q1a3.pdf

[2] BOMFIM, A. 2003. Pre-announcement effects, news effects, and volatility: Monetary policy

and the stock market, Journal of Banking & Finance, volume 27 (issue 1), 133–151. Retrieved

from http://isiarticles.com/bundles/Article/pre/pdf/24857.pdf

[3] BORIO, C., GAMBACORTA, L., HOFMANN B. 2015. The influence of monetary policy on

bank profitability. BIS Working paper No 514. Retrieved from

http://www.bis.org/publ/work514.pdf

[4] FANELLI, P., BACCHIOCCHI, E. 2014. Identification in Structural Vector Autoregressive

models with structural changes, with an application to U.S. monetary policy. Retrieved from

https://www.nbp.pl/badania/konferencje/2014/im/pdf/BACCHIOCCHI_Emanuele.pdf

[5] LIA, H., LIUB, H., SIGANOSB, A., ZHOUC M. 2016. Bank regulation, financial crisis, and

the announcement effects of seasoned equity offerings of US commercial banks. Journal of

Financial Stability, volume 25, 37 – 46. Retrieved from

http://www.sciencedirect.com/science/article/pii/S1572308916300493

[6] MISINA, M., TESSIER, D., DEY, S. 2006. Stress Testing the Corporate Loans Portfolio of

the Canadian Banking Sector.

In Working Paper 2006-47 (ISSN 1192-5434). Retrieved from

http://www.bankofcanada.ca/wp-content/uploads/2010/02/wp06-47.pdf

[7] VIROLAINEN, K. 2004. Macro stress testing with a macroeconomic credit risk model for

Finland. Bank of Finland Discussion Paper No. 18/2004. Retrieved from

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=622682

[8] WANG, P., MOORE, T. 2007. Sudden changes in volatility: The case of five central

European stock markets. Retrieved from

https://www.brunel.ac.uk/__data/assets/pdf_file/0020/84143/07-23.PDF

[9] Board of Governors of the Federal Reserve System. 2015, January 12. Dodd-Frank Act Stress

Test 2015: Supervisory Stress Test Methodology and Results. Retrieved from

http://www.federalreserve.gov/newsevents/press/bcreg/bcreg20150305a1.pdf

[10]Board of Governors of the Federal Reserve System. Press Release. 2015, January 17.

Retrieved from https://www.federalreserve.gov/newsevents/press/bcreg/20150305a.htm

[11]Board of Governors of the Federal Reserve System. 2016, January 12. Dodd-Frank Act

Stress Test 2016: Supervisory Stress Test Methodology and Results. Retrieved from

https://www.federalreserve.gov/newsevents/press/bcreg /bcreg20160623a1.pdf

[12]Board of Governors of the Federal Reserve System. Press Release. 2016, January 17.

Retrieved from https://www.federalreserve.gov/newsevents/press/bcreg/20150305a.htm

[13]European Banking Authority 2014, January 10. Results of 2014 EU‐wide stress test.

Retrieved from http://www.eba.europa.eu/documents/10180/851779/2014%20EU-

wide%20ST-aggregate%20results.pdf

[14]European Banking Authority. 2014, January 15. EBA publishes 2014 EU-wide stress test

results. Retrieved from http://www.eba.europa.eu/-/eba-publishes-2014-eu-wide-stress-test-

results

[15]European Banking Authority 2016, January 10. 2016 EU‐WIDE STRESS TEST RESULTS.

Retrieved from http://www.eba.europa.eu/documents/10180/1532819/2016-EU-wide-stress-

test-Results.pdf

[16]European Banking Authority. 2016, January 15. EBA publishes 2016 EU-wide stress test

results. Retrieved from http://www.eba.europa.eu/-/eba-publishes-2016-eu-wide-stress-test-

results

Kontaktné údaje:

Ing. Lukáš MAJER,

Katedra financií

Národohospodárska fakulta

Ekonomická univerzita v Bratislave

Dolnozemská 1

852 35 Bratislava

Slovenská Republika

majer.lukas@gmail.com

+421 902 271 764

Appendix 1: Bank Shares Volatility analysis